Reference documentation for
The Space Memory — a
cross-workspace knowledge search engine combining FTS5 and vector
retrieval. The CLI is installed as the tsm binary.
tsm is distributed as prebuilt binaries on the
Releases page.
The recommended install path is:
curl -fsSL https://key.github.io/the-space-memory/install.sh | bash
The script picks the right archive for your platform, verifies the
checksum, and copies tsm and tsmd into ~/.local/bin/.
After install, run the one-time setup and the per-workspace initialization:
tsm setup # download embedding model + WordNet DB (system-wide)
cd ~/my-notes
export TSM_INDEX_ROOT=$PWD
tsm init # workspace bootstrap: schema, scaffold, synonym import
tsm start # start the daemon (embedder + file watcher)
tsm index # index your documents
tsm search -q "query" -k 5
For the full quick-start tour, see the project README (日本語版).
| Document | Description |
|---|---|
| Architecture | Daemon / embedder / watcher topology and IPC layout |
| Data Flow | Indexing and search pipelines end-to-end |
| Configuration | tsm.toml and environment variable reference |
| Command Reference | Every tsm subcommand, flags, and examples |
| User Dictionary | Custom terminology for the lindera tokenizer |
| Claude Code prompt format | Output format used by the auto-search hook |